Concrete Pavement Performance in the Southeastern United States

Transcription

1 Concrete Pavement Performance in the Southeastern United States by Norbert J. Delatte, Mudhar Safarjalani, and Natalie B. Zinger Department of Civil and Environmental Engineering The University of Alabama at Birmingham Birmingham, Alabama Prepared by UTCA University Transportation Center for Alabama The University of Alabama, The University of Alabama at Birmingham, and The University of Alabama in Huntsville UTCA Report Number September 2000

2 Technical Report Documentation Page 1. Report No FHWA/CA/OR- 2. Government Accession No. 3. Recipient Catalog No. 4. Title and Subtitle Concrete Pavement Performance in the Southeastern United States 7. Authors Norbert J. Delatte, Mudhar Safarjalani, and Natalie B. Zinger 9. Performing Organization Name and Address Department of Civil and Environmental Engineering The University of Alabama at Birmingham th Street South (Hoehn Building), Suite 120 Birmingham, Alabama Sponsoring Agency Name and Address University Transportation Center for Alabama Box , 271 H M Comer Mineral Industries Building Tuscaloosa, Alabama Report Date September Performing Organization Code 8. Performing Organization Report No. 10. Work Unit No. 11. Contract or Grant No. DTRS98-G Type of Report and Period Covered Final Report/ October 1, July 31, Sponsoring Agency Code 15. Supplementary Notes 16. Abstract This report documents an in-depth study of the performance of concrete pavements in the southeastern United States. Information from the Strategic Highway Research Program (SHRP) Long Term Pavement Performance (LTPP) database was investigated. Analysis of 36 sections in Alabama, Florida, Georgia, Mississippi, and North and South Carolina showed that the majority of these pavements are providing excellent service well beyond their original design lives. This has important implications for new pavement construction, as well as maintenance and rehabilitation of existing pavements. For new pavement construction, the results of this study suggest that life cycle cost models should assume better performance and longer service life than existing AASHTO predictions for these pavements. Thus, the economic benefits of constructing concrete pavements where heavy traffic is anticipated or long life is desired may be considerable. The implication for maintenance and rehabilitation of existing pavements is that concrete pavements may have considerably more remaining structural capacity than time in service or traffic applied to the pavement would suggest. For this reason, expensive and time-consuming reconstruction efforts or thick overlays should not be used unless the evaluation of pavement condition indicates it is warranted. If the pavement is in good structural condition, diamond grinding and other rapid, low cost Concrete Pavement Restoration (CPR) alternatives may extend pavement life considerably and improve serviceability. 17. Key Words Concrete, specifications, quality control, technology transfer, concrete pavements, performance, design, roughness, distress 18. Distribution Statement 19. Security Classif (of this report) 20. Security Classif. (of this page) 21. No of Pages Price Form DOT F (8-72)

3 Contents Contents. iii Tables iv Figures.. v Executive Summary. vi 1.0 Introduction Objectives Scope Plan of Work Organization Background Methodology AASHTO Design Equation Traffic Estimation IRI to PSI Conversion Results and Discussion JPCP Performance Comparison CRCP Performance Comparison. 4-3 Actual and Design Thickness Comparison Conclusions and Recommendations Pavement Design and Rehabilitation Effect of Assumptions Recommendations for Future Research References Acknowledgements Appendix A: Performance Comparison Figures.. A-1 iii

4 List of Tables Number Page 1-1 Pavement sections available and considered Design information from DataPave Design assumptions JPCP sections in study CRCP sections in study Design thickness comparison for JPCP Design thickness comparison for CRCP 4-6 iv

5 List of Figures Number Page 1-1 Road test section , Alabama JPCP, I-59 North, Jefferson County Section , Alabama JPCP, aggregate interlock joints Section , Alabama CRCP A-1 Section , Florida JPCP, doweled joints.... A-1 A-2 Section , Georgia JPCP, doweled joints... A-2 A-3 Section , North Carolina JPCP, doweled joints... A-3 A-4 Section , Mississippi CRCP. A-4 A-5 Section , North Carolina CRCP... A-5 A-6 Section , South Carolina CRCP... A-6 v

6 Executive Summary This report documents an in-depth study of the performance of concrete pavements in the southeastern United States. Information from the Strategic Highway Research Program (SHRP) Long Term Pavement Performance (LTPP) database was investigated. Analysis of 36 sections in Alabama, Florida, Georgia, Mississippi, and North and South Carolina showed that the majority of these pavements are providing excellent service well beyond their original design lives. This has important implications for new pavement construction, as well as maintenance and rehabilitation of existing pavements. For new pavement construction, the results of this study suggest that life cycle cost models should assume better performance and longer service life than existing AASHTO predictions for these pavements. Thus, the economic benefits of constructing concrete pavements where heavy traffic is anticipated or long life is desired may be considerable. The implication for maintenance and rehabilitation of existing pavements is that concrete pavements may have considerably more remaining structural capacity than time in service or traffic applied to the pavement would suggest. For this reason, expensive and time-consuming reconstruction efforts or thick overlays should not be used unless the evaluation of pavement condition indicates it is warranted. If the pavement is in good structural condition, diamond grinding and other rapid, low cost Concrete Pavement Restoration (CPR) alternatives may extend pavement life considerably and improve serviceability. Objectives of this study were: To investigate the performance of concrete pavement in the southeastern United States, as documented in the SHRP LTPP database, To compare performance to that predicted by the AASHTO 1993 design procedure to determine if these sections were meeting or exceeding expectations, To compare actual pavement thickness for each section to design thickness calculated using Portland Cement Association, AASHTO 1993 and AASHTO 1998 procedures to determine which procedure was most appropriate for design in this region, and To determine which design features lead to superior concrete pavement performance in this region. The Strategic Highway Research Program (SHRP) Long Term Pavement Performance (LTPP) database was used to evaluate the performance of some of the 41 concrete pavement sections located in the southeastern United States. Jointed plain concrete pavement (JPCP) and continuously reinforced concrete pavement (CRCP) test sections from the General Pavement Studies (GPS) experiments 3 and 5 were considered, ranging in age from 11 to 34 years. Jointed reinforced concrete pavement (JRCP) was dropped from the analysis because of its limited use in this region, and some other sections were lacking traffic or material information. States vi

7 considered included Alabama, Florida, Georgia, Mississippi, North Carolina, South Carolina, and Tennessee. However, Tennessee sections did not have any information in the database and were dropped from the study. During this time, the growth of population and traffic in the southeast has been explosive. Therefore, it is probable that many or most of these sections are well beyond their initial design traffic volumes. For the 36 sections studied and the assumptions documented in the body of the report, this research suggests the following conclusions and recommendations for concrete pavement design and rehabilitation in the Southeast United States: JPCP with undoweled joints should not be used, particularly for pavements with 5 million or more ESALs of anticipated traffic. Since existing JPCP with undoweled joints may have considerable remaining fatigue capacity, dowel retrofit followed by diamond grinding should be considered to restore serviceability. Doweled JPCP 229 to 254 mm thick (9 to 10 inches) often carries as much as 10 to 20 million ESALs over 25 or more years. The AASHTO 93 procedure appears to produce reasonable designs for doweled JPCP. The AASHTO 98 procedure is unnecessarily conservative for doweled JPCP. CRCP 196 to 234 mm thick (7.7 to 9.2 inches) often carries as much as 7 to 19 million ESALs over 21 to 30 years. Thus, performance similar to doweled JPCP may be achieved with a 10 to 15 % reduction in pavement thickness. These designs may also have considerable remaining life. Both the AASHTO 93 and 98 procedures are unnecessarily conservative for CRCP. Pavements 20 % thinner than required by either AASHTO method have performed very well. Major rehabilitation decisions should be based on pavement condition and not age or cumulative traffic. Pavements 25 to 30 years old that have carried 10 to 25 million ESALs are still performing well, and may be capable of carrying much more traffic. vii

8 Section 1 Introduction Objectives Objectives of this study were: To investigate the performance of concrete pavement in the southeastern United States, as documented in the SHRP LTPP database, To compare performance to that predicted by the AASHTO 1993 design procedure (AASHTO, 1993, Huang, 1993) to determine if these sections were meeting or exceeding expectations, To compare actual pavement thickness for each section to design thickness calculated using Portland Cement Association (PCA, 1984, Huang, 1993), AASHTO 1993 (AASHTO, 1993, Huang, 1993) and AASHTO 1998 (AASHTO 1998) procedures to determine which procedure was most appropriate for design in this region, and To determine which design features lead to superior concrete pavement performance in this region. Scope The Strategic Highway Research Program (SHRP) Long Term Pavement Performance (LTPP) database (FHWA, undated, LAW PCS, 1999) was used to evaluate the performance of some of the 41 concrete pavement sections located in the southeastern United States. Jointed plain (JPCP) and continuously reinforced (CRCP) test sections from the General Pavement Studies (GPS) experiments 3 and 5 were considered, ranging in age from 11 to 34 years. States considered included Alabama, Florida, Georgia, Mississippi, North Carolina, South Carolina, and Tennessee. However, Tennessee sections did not have any information in the database and were dropped from the study. During this time the growth of population and traffic in the southeast has been explosive. Therefore, it is probable that many or most of these sections are well beyond their initial design traffic volumes. The pavement designs available and considered by type and state are listed in Table 1-1. Three jointed reinforced concrete pavement (JRCP) sections (GPS 4) were found in the database but were not considered further in this study. The limited sample suggests that this type of pavement is not used much in the region. Two sections, both JPCP, did not have enough data available and could not be studied further. Either traffic or material information was insufficient. Thus, the final study sample comprised 22 JPCP and 14 CRCP sections. Twelve sections from the PCC structural factors (SPS-2) experiment in North Carolina were also investigated. However, these sections are only about 5 years old and no significant distress has been observed, so it is too early to analyze them. A sign showing the location of a typical section is shown in Figure

9 Table 1-1. Pavement Sections Available and Considered State JPCP JRCP CRCP Total Available Considered Available Considered Available Considered Available Considered Alabama Florida Georgia Mississippi North Carolina South Carolina Tennessee Total Figure 1-1. Road test section , Alabama JPCP, I-59 North, Jefferson County. Plan of Work The following tasks constituted the work plan for this project: a. Conduct a literature review, with special attention to the Federal Highways Administration reports documenting the experimental design and data collection methods and lessons learned on concrete pavement performance. b. Design an investigation using the database to determine the design features that improve pavement performance in this geographic region. c. Retrieve and analyze the data from the database. 1-2

10 d. Analyze selected sections using the AASHTO 1993 equations, the PCA method, and the new 1998 AASHTO procedure to determine whether the performance has been better or worse than predicted by design procedures currently in use. e. Write a final report documenting the results of the study. Organization The experimental design and results are presented in the following sections. Background is presented in Section 2 and methodology in Section 3. Section 4 presents project results and discussion, and conclusions and recommendations are provided in Section

11 Section 2 Background For the 36 sections considered, the first step was to develop comparisons between actual and predicted performance and analyze the results. Next, the actual pavement thickness for each section was compared to the design thickness obtained using a variety of procedures. This methodology is discussed in section 3. For the performance comparison, measured pavement International Roughness Index (IRI) was converted to Present Serviceability Index (PSI) and plotted against the 1993 AASHTO design equation. Some design information could be extracted directly from the database using the DataPave 2.0 software, as listed in Table 2-1. Table 2-1. Design Information from DataPave Design Parameter Pavement thickness Pavement roughness Traffic, in ESALs Soil under pavement Shoulder type Drainage Joint type, spacing DataPave Information Mean thickness from section report International Roughness Index (IRI) by year Estimated and measured ESALs, by year (not complete) California Bearing Ratio (CBR) or AASHTO Soil Classification Tied (concrete) or asphalt Undrained, longitudinal, drainage blanket Doweled, aggregate interlock The data fields listed in DataPave 2.0 appear at first glance to provide all necessary information for analysis and design. However, on closer inspection, many data are missing and must be assumed from other parameters. For example, the modulus of subgrade reaction k, an important design input, was not provided for the sections considered. Table 2-2 lists the assumptions made as well as the basis for those assumptions. Table 2-2. Design Assumptions Design Assumption Basis Total cumulative traffic Estimated and measured traffic and interpolation Modulus of subgrade reaction k Correlations to CBR and AASHTO Soil Classification (PCA 1984, AASHTO 1998) Concrete flexural strength Assumed 4.48 MPa (650 psi) (AASHTO 4993) Concrete modulus of elasticity Assumed 25.6 Gpa (3.71 million psi), consistent with assumed concrete flexural strength Standard deviation Assumed 0.34 based on AASHTO (AASHTO 1993), traffic variation not considered Reliability 50 % for comparing performance (Figures 1 through 8) and 85 % for determining design requirements (Tables 6 and 7) Drainage coefficient C d 0.9 to 1.1, based on drainage type Joint J 2.6, 2.8, 3.1, 3.2, or 4.1 based on pavement, joint, and shoulder type Data on AASHTO soil classification and California Bearing Ratio (CBR) is available in many cases, and k may then be estimated using Portland Cement Association (PCA, 1984, Huang, 1993) or AASHTO supplement (AASHTO 1998) charts. In the absence of other data, the layer descriptions provided in the Section Reports may be used for rough estimates along with 2-1

12 AASHTO guide charts (AASHTO 1986, AASHTO 1993). In some cases, curves representing the highest and lowest estimated k values were developed and shown on the same figure. For Alabama and Mississippi, the CBR could be used to find a single value of k. For other states, the AASHTO soil classification was used, which may be correlated to a range of k values. Using this design information and an assumed initial serviceability index of 4.5, traffic in ESALs corresponding to PSI values of 4.0 to 0 (at 0.5 increments) was calculated and plotted as shown in the figures in Section 4. Some figures represent cases where two different values of k were used. Next, field measured IRI values converted to PSI were plotted, using an equation from Hall and Correa (1999). Although 36 plots were generated during this research, only 8 representative figures are provided in this report. The cumulative traffic corresponding to the measured IRI values was projected using both estimated and measured traffic by year from the database. Although this information was incomplete and in some cases contradictory, the cumulative traffic is probably reasonable since overestimation and underestimation would tend to cancel out over the 11 to 34 year period. Where both measured and reported traffic were shown for the same year, they often did not agree one or the other was used, based on which better matched the traffic information from previous and subsequent years. Cumulative traffic ranged from less than half a million to nearly 30 million 80 kn (18 kip) Equivalent Single Axle Loads (ESALs). 2-2

13 Section 3 Methodology AASHTO Design Equation The 1993 AASHTO design equation was used to compare design performance under predicted traffic load with actual performance AASHTO rigid pavement design equation (AASHTO 1993) is: log 10 ( ) = + ( + ) + W 18 ( ) 4.22 Z R S O p t log log D 215 S 1 / C.63 C J 0.06 d 0.75 ( ) D D log E c PSI ( D + 1 ) k 0.25 ( / ) 7 ( 3 1) Where W 18 Z R S O D predicted number of 80 kn (18-kip) Equivalent Single Axle Load (ESAL) applications determined from the equation above and then converted to thousand of a Single Axle Load (KESAL). standard normal deviate assumed to be equal to 0 for reliability of 50% ("actual" performance) and to for reliability of 85% (design performance), where higher reliability requirement demands a thicker design. combined standard error of the traffic prediction and performance prediction assumed to be equal to 0.35 for rigid pavements. thickness (25 mm increments or inches) of pavement slab provided by the database. PSI difference between the initial design serviceability index, p o, and the design terminal serviceability index, p t, which varies from 0 (for p o = p t = 4.5) to 4.5 (for p o = 4.5, p t = 0) S c modulus of rupture (psi) for Portland cement concrete used on a specific project assumed to be equal to 4.48 MPa (650 psi ) for concrete with normal aggregates. J load transfer coefficient used to adjust for load transfer characteristics of a specific design. The material type (Asphalt or Tied PCC) and presence of Load Transfer Devices 3-1

14 for outer shoulder of each section were used to make assumptions. The range of values used was C d drainage coefficient. For the Pavement s Exposure to Saturation from 5 to 25% applicable to the region of interest, C d was assumed to be equal to 1.10 for drainage blanket and longitudinal drains, 1.00 in case of unknown drainage systems, and 0.90 for no drainage system. E c k modulus of elasticity (psi) for Portland Cement Concrete (PCC). modulus of sub-grade reaction (psi) was listed as a field in DataPave 2.0. However, no k values were available for the sections of interest. Therefore, k was estimated from California Bearing Ratio and/or from the AASHTO type of soil and adjusted according to the treatment of section's sub-layers. Assuming that initial serviceability level of the constructed section was 4.5 (PSI) and eventually decreased, the interval of serviceability was 4.5 to 0.0. To represent graphically a design model, an increment of serviceability decrease of 0.5 was used. That gave 10 design points to generate a curve. It was not known which Standard Normal Deviate (Z R ) was considered in the design procedure. Therefore, two values of Z R were involved in analysis of the design model - one for reliability of 50% and another for 85%. This difference in Z R values was reflected in the difference in slopes of graphs representing pavement performance. Another uncertainty in the presentation of the design curve was dependant on the choice of a k value. For some states (Alabama and Mississippi) a CBR value was found in LTPP database. Other states information was interpreted in terms of AASHTO soil classification. By converting a CBR value into modulus of sub-grade reaction, one gets a single value, while AASHTO soil type provides a range of k values for a given soil type (AASHTO 1998). Since there was no method known to determine which modulus from this range was most descriptive for the section of interest, it was logical to use two extremes of this range. Therefore, two values of modulus of sub-grade reaction appear in calculations: k max and k min. Actual pavement performance would be expected to fall in-between of two design curves based on two k values. Assumptions made above resulted in generating of 2 to 4 design curves of pavement performance. These design curves were analyzed and compared with actual pavement performance. Because two different k values could be assumed, five of the figures in Appendix A (all but A-4) have two AASHTO curves. In this case the two curves should be considered as the upper and lower limits on a band. Traffic Estimation In order to compare the actual traffic data provided by LTPP database with the results of the AASHTO design procedure, it was necessary to present these two sets of data in compatible formats. Since the AASHTO equation is based on the relationship between traffic in the Equivalent Single Axle Load (ESAL) and level of serviceability in Present Serviceability Index (PSI), it was necessary to make sure that LTPP provided equivalent actual data for a section of interest. The program s database has two types of traffic information: estimated and monitored. This information is stored in DataPave 2.0 under TRF_EST_ANL_TOT_LTPP_LN and 3-2

15 TRF_MONITOR_BASIC_INFO respectively. These were used to estimate a cumulative traffic in ESALs. IRI to PSI Conversion The serviceability level of a section is described in the database through the International Roughness Index (IRI), so it had to be converted to PSI in order to be compared with AASHTO design model. The IRI value was converted to PSI using a relationship from Hall and Correa (Hall and Correa 1999). Where and PSI = x x x (3-2) x = log (1 + SV) (3-3) SV = IRI 2 (3-4) It was established that FHWA LTPP program database presents required data in terms of IRI and two types of traffic data: estimated and monitored. It is also known that AASHTO design model procedure applied to design pavements represents the same data in terms of PSI and predicted traffic load. Actual performance information was limited by the number of times the roughness of a section was measured. Another factor limiting the analysis was the lack of measured traffic data. A couple of readings were required in order to estimate the value of traffic for a certain year a year for which PSI was given or any other year. The estimation of traffic for that particular year had to be performed through interpolation and/or extrapolation of given traffic data. A value of the cumulative traffic for the year of interest was a summation of loads for all the years a section was in use. Although some gaps were found in LTPP database, they were filled with estimated values of section s traffic. To create a graph at least two points were needed. Therefore, if it was impossible to find at least two values for actual serviceability of a section and to closely estimate a traffic load for the same years, it was not possible to generate a graph of actual pavement performance. For those sections there was nothing to compare with design graphs based on the AASHTO design model and no analysis of pavement performance could be done. As a result, 4 of the 42 sections could not be analyzed further: , , , and Following the analytical procedure described in this section, a graph was created for each section included in the study. 3-3

16 Section 4 Results and Discussion Actual performance versus predicted performance was plotted for the 36 sections for which enough design and traffic data was available. Examples are shown in Figures 4-1 and 4-2, and in Appendix A in Figures A-1 through A-6. For the AASHTO performance prediction, the mean (50 % reliability) was assumed PSI ,000 2,000 3,000 4,000 5,000 6,000 KESAL AASHTO Figure 4-1. Section , Alabama JPCP, aggregate interlock joints. JPCP Performance Comparison Sample results for JPCP sections are shown in Figure 4-1 and in Appendix A, Figures A-1 through A-3. Information on the 22 JPCP sections is provided in Table 4-1. The table includes state, section number, and route, age (as of the year 2000), drainage type, joint type, IRI (as of the year 1999), PSI (estimated from IRI), Average Daily Truck Traffic (ADTT, provided for the latest year available), and cumulative traffic in thousands of ESALs (KESALs). JPCP sections ranged in age from 11 to 34 years, were 201 to 284 mm (6.4 to 13.3 inches) thick, and had cumulative traffic of 388 thousand to 27.5 million ESALs. 4-1

17 Table 4-1. JPCP Sections in Study State Age Slab Type of Type of IRI PSI ADTT Cumul. Section # Route # (years) thickness drainage joint (m/km) (est.) KESALs Alabama mm inches blanket aggregate Florida none dowels long. drains aggregate blanket dowels long. drains dowels long. drains dowels none aggregate Georgia none dowels blanket none long. drains dowels long. drains dowels blanket aggregate blanket aggregate none dowels none dowels Mississippi none dowels none dowels North Carolina none dowels unknown dowels unknown dowels none none none dowels South Carolina none dowels Tennessee None Alabama section (Figure 4-1) represents the worst performance in the study, with a current PSI of only about 2.0 after 5.5 million ESALs. Closer inspection of the damage mechanisms reveals considerable joint faulting, which is to be expected since the pavement had aggregate interlock joints. Other sections without doweled joints, such as Florida sections and , also performed poorly. These are probably good candidates for dowel retrofit followed by diamond grinding. In contrast, other sections studied showed much better performance, as documented in Appendix A. Florida section (Figure A-1) is in excellent condition after more than 11 million 4-2

18 ESALs. At 13.3 inches, it is the thickest pavement in the study. It is also one of the newest pavements in the study sample, with a construction date of Georgia section (Figure A-2) also has excellent serviceability after 18.8 millions ESALs and 27 years of service. North Carolina section (Figure A-3) has seen the heaviest traffic, 27.5 million ESALs in 23 years, yet still has a serviceability index near 3.5. Its current serviceability is more than twice as high as predicted by the 1993 AASHTO equations. Overall, of the 22 JPCP sections, all but 3 have a current PSI greater than 3.0, as estimated from IRI. Those three all have aggregate interlock rather than doweled joints. At 3.0 level of serviceability, less than 12 % of users would see a need for rehabilitation or other action (AASHTO 1993). With the exception of Florida, these pavements are 201 to 284 mm (7.9 to 11.2 inches) thick. Florida sections ranged from 163 to 338 mm (6.4 to 13.3 inches), a much wider range than in other states. Most of the performance comparison plots for JPCP gave reasonably good agreement between the AASHTO 93 equation prediction and observed field performance, generally within 0.5 to 1 PSI. CRCP Performance Comparison Sample results for CRCP sections are shown in Figure 4-2 and in Appendix A, Figures A-4 through A-6. Information on the 14 CRCP sections is provided in table 4-2. CRCP sections ranged in age from 21 to 30 years, were 196 to 234 mm (7.7 to 9.2 inches) thick, and had cumulative traffic of 2.44 to 22.2 million ESALs. Thus, the variations in age, pavement thickness, and traffic are much smaller for the CRCP sections than the JPCP sections. Alabama section (Figure 4-2) represents excellent performance, with a current PSI of over 4.5 after 11.4 million ESALs and 24 years. Similar trends are seen for other sections, as represented in Appendix A by Mississippi section (Figure A-4) and North Carolina section (Figure A-5), both of which have serviceability above 4.0 after 25 to 27 years and 17.7 and 3.24 million ESALs, respectively. Section (Figure A-6) from South Carolina represents the only CRCP section that is performing below expectations, and the difference between actual and predicted performance is only about 0.5 PSI. Overall, all 14 CRCP sections have a current PSI greater than 3.33, and all but three are greater than 4.0. In all but two cases, actual performance exceeds the AASHTO 93 equation prediction, often by a considerable margin. Figures 4-1, A-4, and A-5 are representative of the 12 sections that are exceeding expectations. Actual and Design Thickness Comparison The actual pavement thickness for each section is compared to the PCA, AASHTO 1993, and AASHTO 1998 design thickness in Table 4-3 for JPCP and Table 4-4 for CRCP sections. Only sections with heavy traffic (more than 5 million ESALs) were considered. The design thickness in each case was calculated to carry the traffic that had already been imposed on the pavement (5 to 27.5 million ESALs over 11 to 34 years), to a current PSI of 3.0. Note that unlike the performance plots, which did not use a reliability term, a reliability of 85 % was assumed for the AASHTO 93 and 98 procedure design calculations. 4-3

19 PSI ,000 4,000 6,000 8,000 10,000 12,000 14,000 KESAL AASHTO Figure 4-2. Section , Alabama CRCP For example, Alabama section has an actual thickness of 259 mm (10.2 inches). For an ADTT of 2106, the PCA design procedure predicts a thickness of 221 mm (8.7 inches), the AASHTO 98 procedure a thickness of 250 mm (9.83 inches) and the AASHTO 93 procedure a thickness of 252 mm (9.91 inches). Dividing the design values by the actual thickness gives ratios of 0.85, 0.96, and 0.97 respectively between the PCA, AASHTO 98, and AASHTO 93 designs and actual thickness, as shown in Table 6. For cases where the AASHTO soil classification was used to predict k, minimum and maximum design thicknesses corresponding to maximum and minimum k were developed. In most cases the ratio for the PCA method is low. There are several possible explanations. The ADTT used may not be consistent with the cumulative number of ESALs. Also, the PCA design method uses fatigue and erosion failure criteria, which would be consistent with a PSI much less than 3.0. As a result, it is probably not possible to draw conclusions from the PCA comparison. In contrast, the AASHTO 98 designs exceed actual thickness by 8 to 27 % for JPCP, and 20 to 28 % for CRCP, on average. AASHTO 93 designs are closer to actual thickness for JCPC (3 % less to 5 % more, on average), but 19 to 22 % high for CRCP, on average. 4-4

20 Table 4-2. CRCP Sections in Study State Age Slab Type of IRI PSI ADTT Cumul. Section # Route # (years) thickness drainage (m/km) (est.) KESALs Alabama mm inches None None none Florida Georgia None Mississippi None None None None None North Carolina Unknown None Unknown South Carolina None None None Tennessee none Table 4-3. Design thickness comparison for JPCP PCA AASHTO 98 AASHTO 93 Section min max Min max min max AL FL FL FL GA GA GA NC NC SC Averages: Standard Deviation: For JPCP, the AASHTO 93 designs are thinner than the actual pavement for only three sections, if both minimum and maximum values for k are considered. Only two of these would also have thinner designs under the AASHTO 98 Supplement procedure. 4-5

21 Table 4-4. Design thickness comparison for CRCP PCA AASHTO 98 AASHTO 93 Section min max Min max min max AL AL GA MS MS MS NC NC SC SC SC Averages: Standard Deviation:

22 Section 5 Conclusions and Recommendations Pavement Design and Rehabilitation For the 36 sections studied and the assumptions documented in the body of the report, this research suggests the following conclusions and recommendations for concrete pavement design and rehabilitation in the Southeast United States: JPCP with undoweled joints should not be used, particularly for pavements with 5 or more million ESALs of anticipated traffic. Since existing JPCP with undoweled joints may have considerable remaining fatigue capacity, dowel retrofit followed by diamond grinding should be considered to restore serviceability. Doweled JPCP 229 to 254 mm thick (9 to 10 inches) often carries as much as 10 to 20 million ESALs over 25 or more years. The AASHTO 93 procedure appears to produce reasonable designs for doweled JPCP. The AASHTO 98 procedure is unnecessarily conservative for doweled JPCP. CRCP 196 to 234 mm thick (7.7 to 9.2 inches) often carries as much as 7 to 19 million ESALs over 21 to 30 years. Thus, performance similar to doweled JPCP may be achieved with a 10 to 15 % reduction in pavement thickness. These designs may also have considerable remaining life. Both the AASHTO 93 and 98 procedures are unnecessarily conservative for CRCP. Pavements 20 % thinner than required by either AASHTO method have performed very well. Major rehabilitation decisions should be based on pavement condition and not age or cumulative traffic. Pavements 25 to 30 years old that have carried 10 to 25 million ESALs are still performing well, and may be capable of carrying much more traffic. This report documents an in-depth study of the performance of concrete pavements in the southeastern United States. Information from the Strategic Highway Research Program (SHRP) Long Term Pavement Performance (LTPP) database was investigated. Analysis of 36 sections in Alabama, Florida, Georgia, Mississippi, and North and South Carolina showed that the majority of these pavements are providing excellent service well beyond their original design lives. This has important implications for new pavement construction, as well as maintenance and rehabilitation of existing pavements. For new pavement construction, the results of this study suggest that life cycle cost models should assume better performance and longer service life than existing AASHTO predictions for these pavements. Thus, the economic benefits of constructing concrete pavements where heavy traffic is anticipated or long life is desired may be considerable. 5-1

23 The implication for maintenance and rehabilitation of existing pavements is that concrete pavements may have considerably more remaining structural capacity than time in service or traffic applied to the pavement would suggest. For this reason, expensive and time-consuming reconstruction efforts or thick overlays should not be used unless the evaluation of pavement condition indicates it is warranted. If the pavement is in good structural condition, diamond grinding and other rapid, low cost Concrete Pavement Restoration (CPR) alternatives may extend pavement life considerably and improve serviceability. Effect of Assumptions The AASHTO 1993 design equation (Equation 3-1) is more sensitive to some variables than others. The assumptions made in this study are reasonable and consistent with the available data. Modulus of subgrade reaction k is one of the more difficult parameters to estimate, but has little effect on the performance prediction. This may be seen in the figures in Appendix A the two curves are close together even though the higher k values are four times the lower. Overall, the performance prediction is relatively insensitive to material parameters (k, S c,, and E c ). The model is more sensitive to load transfer and drainage coefficients (J and C d ) but these can be estimated accurately with the information from DataPave. Recommendations for Future Research The results of this study suggest a number of potential avenues for further work: The methodology of this study could easily be extended to other regions, or to the entire United States. This could be used to evaluate the performance of JPCP and CRCP in a variety of climate regimes, and not merely the wet-no freeze regime prevalent in the southeast. These results could provide valuable input for the development of the 2002 AASHTO Pavement Design Guide and related documents. Alternatively, it would be possible to pursue this research in greater depth, using nondestructive testing and forensic analysis. Regional and state-specific analyses could be carried out, with core testing for concrete properties and study of fallingweight deflectometer results. There are many other topics in the area of pavement performance that could be investigated, using the expertise that UAB has acquired with DataPave

24 Section 6 References American Association of State Highway and Transportation Officials (AASHTO). AASHTO Guide for Design of Pavement Structure, Volume 2. Washington, DC American Association of State Highway and Transportation Officials (AASHTO). Guide for Design of Pavement Structures. Washington, DC American Association of State Highway and Transportation Officials (AASHTO). Supplement to the Guide for Design of Pavement Structures. Washington, DC Federal Highway Administration (FHWA). Long Term Pavement Performance DataPave Demonstration Workshop Participant s Notebook. Undated. Hall, K. T., and Correa Muñoz, C. Estimation of Present Serviceability Index from International Roughness Index, paper no presented at the 78 th Annual Meeting of the Transportation Research Board, Washington, D.C Huang, Y. H. Pavement Analysis and Design. Prentice-Hall, Englewood Cliffs, N.J LAW PCS, a LAWGIBB Group Member. Introduction to LTPP Data. Office on Infrastructure Research and Development, US DOT, FHWA, McLean, VA Portland Cement Association (PCA). Thickness Design for Concrete Highway and Street Pavements. Skokie, IL

25 Section 7 Acknowledgements Funding for this research was provided by the University Transportation Center for Alabama, the Southeast Chapter of the American Concrete Pavement Association, and the University of Alabama at Birmingham, and their support was much appreciated. However, opinions expressed in this paper are those of the authors alone and not necessarily of those organizations. 7-1

26 Appendix A Performance Comparison Figures Performance comparison figures for six sections are provided below. In cases where two AASHTO lines are provided, they represent higher and lower values of k, respectively, and should be considered as the upper and lower limits on a band PSI ,000 4,000 6,000 8,000 10,000 12,000 KESAL AASHTO Figure A-1. Section , Florida JPCP, doweled joints. A- 1

27 PSI ,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 KESAL AASHTO Figure A-2. Section , Georgia JPCP, doweled joints. A- 2

28 PSI ,000 10,000 15,000 20,000 25,000 30,000 KESAL AASHTO Figure A-3. Section , North Carolina JPCP, doweled joints. A- 3

29 PSI ,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 KESAL AASHTO Figure A-4. Section , Mississippi CRCP. A- 4

30 PSI ,000 1,500 2,000 2,500 3,000 3,500 KESAL AASHTO Figure A-5. Section , North Carolina CRCP. A- 5

31 PSI ,000 2,000 3,000 4,000 5,000 6,000 7,000 KESAL AASHTO Figure A-6. Section , South Carolina CRCP. A- 6